Dworkin et al.
/
5299
Reactions of Arenes in Molten Antimony Trichloride
(17) (a) J. W. Timberlake and J. C. Stowell in "The Chemistry of the Hydrazo, Azo and Azoxy Groups", S.Patai, Ed., Wiley-lnterscience, New York, 1975, pp 69-108; (b)J. W. Timberlake,M. L. Hodges. and K. Betterton, Synthesis,
632 (1972). (18) R. F. Weinland and W. Stille, Chem. Ber., 34, 2631 (1901). (19) D. Naumann, L. Deneken, and E. Renk. J. Fluorine Chem., 5, 509 (1975). (20) (a)E. E. Aynsley, R. Nichols, and P. L. Robinson, J. Chem. Soc., 623 (1953);
(b) J. W. Viers and H. W. Baird, Chem. Commun., 1093 (1967);(c) A. J. Edwards and P. Taylor, J. Fluorine Chem.. 4, 173 (1974). (21) C. Hartmann and V . Meyer, Chem. Ber.. 26, 1727 (1893). (22) See. for example, A. Engelbrecht. P. Peterfy, and E. Schandara, Z.Anorg. Allg. Chem.. 384, 202 (1971). (23) R. L. Ameyand J. C. Martin, J. Org. Chem., 44, 1779(1979). (24) C. H. Dungan and J. H. van Wazer, "Compilation of Reported F19 NMR
Chemical Shifts", Wiley-lnterscience,New York, 1970.
Hydrogen Transfer Reactions of Arenes in Molten Antimony Trichloride A. S. Dworkin,* M. L. Poutsma, J. Brynestad, L. L. Brown, L. 0. Gilpatrick, and G. P. Smith Contribution from the Chemistry Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830. Received March 21, 1979
Abstract: The reactions of polycyclic aromatic hydrocarbons in molten antimony trichloride at temperatures from 80 to 175 "C have been studied by in situ IH N M R and quench and separation techniques. Decomposition takes place by a complex series of hydrogen redistribution reactions whose net effect is a disproportionation of a type not previously known to occur. The reaction of anthracene ( I ) was studied in detail. Some of the anthracene molecules lose aryl-bound hydrogen and are condensed into larger molecules with aryl-aryl bonds such as the asymmetric bianthracenes (IV, V , and V I ) and anthraaceanthrylenes (VI1 and VIII). The hydrogen released by these reactions is quantitatively captured by other anthracene molecules to form hydroaromatic molecules such as the 9,lO-dihydro- and 1,2,3,4-tetrahydroanthracenes(I1and I l l ) . The catalytic role of the solvent's Lewis acidity was demonstrated by the fact that the addition of a few mol % of a strong chloride donor reduced the reaction rates by a large factor. The reaction behavior of a group of related arenes (naphthalene, phenanthrene, chrysene, pyrene, perylene, and naphthacene) of widely varying basicity and oxidizability was also surveyed.
Introduction We have briefly reported' that anthracene (I) dissolved in high-purity molten antimony trichloride under an inert atmosphere undergoes spontaneous hydrogen transfer reactions at temperatures as low as 100 OC with I serving as an efficient hydrogen acceptor for its own Scholl condensation reactions.2 It has long been known that strong Lewis acids, such as molten aluminum trichloride, catalyze dehydrogenation-condensation reactions of aromatic compounds (Scholl reactions), but the fate of the hydrogen thus released had not been demonstrated. In the absence of added oxidizing agents, it has generally been assumed that hydrogen is evolved as a gas. Thus, our results are novel in two important respects: first, they demonstrate that the substrate aromatic can serve as an efficient hydrogen acceptor for its own Scholl reactions; and second, they demonstrate that Scholl reactions can be catalyzed by a mild Lewis acid, antimony trichloride. A full account of this aspect of Friedel-Crafts chemistry is reported in this article. The reaction behavior of a group of related arenes with widely varying basicity and oxidizability is also described. Most prior research on the behavior of polycyclic aromatics in liquid antimony trichloride was done at 74-99 OC3 where we found that anthracene disproportionates quite slowly. Some research at higher temperatures4 encountered reactions of an undetermined nature. Results and Discussion Solutions of anthracene in SbC13 prepared by melting together the high-purity components at 80 OC under an inert argon atmosphere were initially yellow. After about 2 h at this temperature, the solutions took on a greenish tinge. This very slow color change was greatly accelerated by increasing temperature such that at 125 OC the solutions were a very dark green almost immediately. Except for the first few minutes at
0002-78631791 I501-5299$01 .OO/O
125 OC, the 100-MHz ' H N M R spectrum of the melt was dominated by a pattern fully consistent in position and fine structure with that of molecular anthracene in nonreactive solvents (CDC13). The appropriateness of antimony trichloride as an N M R solvent was demonstrated by the observation of the spectrum of the triphenylmethyl cation comparable to that reported in FS03H.S The bands in the anthracene spectrum were initially broadened with the bands of Hg.10 being the broadest and those for H2.3.6.7 being narrowest at any given instant. Since this order also corresponds to the relative magnitude of the hyperfine coupling constants in the anthracene radical cation, the broadening was attributed to exchange with traces of radical cation.6 However, the spectrum rapidly sharpened as the radical cations decayed or as the unpaired electron was transferred to products of greater oxidizability. The presence and decay of these radical cations were confirmed by ESR at 80 OC where the decay rate was much slower. Both the color change and spectral sharpening were irreversible since they persisted when the solution was cycled between 125 and 80 OC. Coincident with the sharpening of the anthracene spectrum, the lines for 9,lO-dihydroanthracene (11) were observed and 6 7
m m a -.. -..
5
\ 23 /
(0 4
I
I1
I11
grew with time at the expense of those of anthracene. The methylene singlet at 6 = 3.99 ppm (referenced to Me4Si = 0) and ten resolvable lines of the AA'BB' pattern centered at 6 = 7.33 ppm for I 1 were completely distinguishable from the anthracene spectrum in which no lines were further upfield than 6 = 7.5 ppm. After about an hour at 125 "C, bands at 6 = 1.85 ppm and 2.93 ppm attributable to the methylene groups i n 1,2,3,4-tetrahydroanthracene(111) appeared and also continued to grow with time. Hydrogenated products I1 and 111
0 I979 American Chemical Society
Journal of the American Chemical Society / 101:18 / August 29, 1979
5300
'0°
-
r ANTHRACENE
E I
I
-111
0
50
400
150 0 50 400 R E A C T I O N TIME (minl
450 200
Figure 1. Effect of the initial concentration of anthracene and the temperature on the rates of formation of 9,lO-dihydroanthracene (11) and 1,2,3,4-tetrahydroanthracene(111) in molten SbC13. The weight of total organics (equal to the initial weight of anthracene) remains invariant throughout the experiment.
were initially identified by comparison with the 'H N M R spectra of the authentic compounds in molten SbC13. The spectra were well-resolved and very similar to the corresponding spectra for solutions in CDC13. The N M R identification was confirmed by H P L C and G C separations and UV spectra. The reaction rates were found to be independent of the initial concentration of I over the range of approximately 2-10 mol 70,whereas a n increase in temperature of less than 25 O C increased the rates by a large factor. Both of these effects are illustrated in Figure 1 which is based on integration of the N M R spectra taken as the reactions proceeded. When I1 is dissolved in antimony trichloride a t 130 OC, it disproportionated to I and 111(and the products of the reaction of I) at a rate slightly less than that shown in Figure 1 for formation of I1 from 1. The N M R spectrum of 111 dissolved in antimony trichloride was stable for a few hours after which some change in the spectrum was seen but not the formation of either I or 11. The above reaction behavior is consistent with that shown in Figure 1; Le., a maximum of the amount of I1 was reached after above 2-h of reaction time followed by its decrease and the continuous increase of the amount of 111. Under the conditions of the experiment adventitious hydrogen could not possibly have accounted for the degree of reaction described above. Since no other external hydrogen source was present, the formation of hydro derivatives of anthracene implies the formation of some coproducts containing less hydrogen than anthracene which suggests, therefore, the coupling of anthracene units. Although these condensed products were not seen as resolved entities in the N M R spectra, extra absorption appeared as a broad nondescript envelope lying under the distinct anthracene lines and grew as I decreased. In order to identify the condensed products and also to confirm the N M R integration results for the hydrogenated products, quench and separation techniques were employed. The components of the organic phase recovered after a run were separated principally by HPLC and T L C and identified by UV, N M R , mass spec, and melting point. We found by H P L C that the condensed aromatic material was comprised of approximately 20 different compounds. However, a significant fraction of the condensed material during the first hour or two of the reaction was made up of just two of these compounds, an asymmetric bianthracene (B) which is either the 1,9'-bianthracene (IV) or the 2,9'-bianthracene ( V ) (see Experimental Section), and anthra[2, I-alaceanthrylene (VII). Evidence from UV and mass spectra suggested that the two other asymmetric bianthracenes (VI and either IV or V ) , anthra[ 1,2-a]aceanthrylene (VIII), and dibenzo[a,j]perylene
CONDENSED OLIGOANTHRACENES
TIME (mini
Figure 2. Rates of disappearance of anthracene and formation of the foui principal products at 130 O C ; initial concentration, 10 mol 70anthracene in SbCI3. The weight of total organics (equal to the initial weight of anthracene) remains invariant throughout the experiment.
VI1
vu1
Ix
(IX) were among the minor products formed. The remainder was composed of more condensed materials (oligoanthracenes formed from two to six anthracene units have been identified by mass spec) and other unidentified substances. There are six possible isomers of bianthracene; the preparation and properties of the three symmetrical isomers are reported in the l i t e r a t ~ r e ' .but ~ those of the asymmetric ones have not been reported. The selectivity for asymmetric bianthracenes when I is heated a t moderate temperatures for relatively short periods of time in molten SbC13 suggests that new synthetic techniques may be derived from this work. Figure 2 summarizes the course of the reactions for the first 400 min as monitored by both N M R and separation techniques. The principal asymmetric bianthracene (B) formed rapidly coincident with I I and appeared to be the precursor of VI1 which decayed more slowly than did B. In order to confirm that VI1 actually was derived from B, we studied the reactions of VI1 and of the principal bianthracene (B) in molten SbC13. After 1 h a t 130 OC, only 5% of the bianthracene remained. The main condensation product was found to be VI1 (about 40% of the total organics) and the reduced products identified were 1, 11, and 111. Again, a number of products were found which could not be identified. On the other hand, more than 90% of the original amount of VI1 was recovered from a solu-
Dworkin et al.
/
Reactions of Arenes in Molten Antimony Trichloride
tion of VI1 in molten SbC13 kept at 130 "C for 1 h. It is of note that authentic 9,9'-bianthracene was found to be stable for an extended period of time in molten SbC13 a t 130 O C and yet we found only trace amounts of it among the condensed products of the reaction of I in SbC13. In order to ascertain the extent to which the hydrogen liberated from the condensation reactions was transferred to hydrogenated compounds in the molten reaction medium, the gas phase above the melt was also analyzed. After a typical experiment the reaction tube was cooled, connected to a mass spectrometer, and opened to the spectrometer through a break seal. N o hydrogen-containing gases (Hz, HCI, or organics) in amounts greater than 0.001% of the hydrogen in the original anthracene reactant were found. This evidence, then, reconfirms the efficiency of the hydrogen transfer in the molten antimony trichloride reaction medium. The addition of a few mole percent of a strong chloride donor, KCI, or (CH&NCI greatly reduced the rate of the anthracene disproportionation reaction. For example, in the presence of 3 mol % (CH3)4NCI after 1500 min a t 130 "C, I1 made up 14% of the organics whereas less than 40 min were required to form an equivalent amount in the absence of (CH3)4NCI. At the same time (1 500 min) with added chloride ion donor, I11 was not as yet discernible in the N M R spectrum. The maximum yield of I1 with added chloride ion donor occurred a t about 2400 min as compared with about 120 min in the neat melt while 111 comprised about 10%of the organics a t 2400 min, a concentration reached in about 180 min in the neat melt. The addition of a chloride ion donor slowed the rate of disproportionation of I1 in SbCI3 in a like manner. This strong effect on the reaction rate suggests that the solvent's Lewis acidity is playing a catalytic role. SbC13 is very weakly ionized to SbC12+ (a moderately strong Lewis acid) and chloride ion (solvated in some manner) with an ion product constant variously reported from 1.6 to to
